UntargetedAttack

class robustcheck.types.UntargetedAttack.UntargetedAttack(model, img, label)

Abstract class for untargeted adversarial attacks

This is an abstract class that provides a template for standard untargeted adversarial attacks.

model

Target model to be attacked. This has to expose a predict method that returns the output probability distributions when provided a batch of images as input.

img

An array (HxWxC) representing the target image to be perturbed.

label

An integer representing the correct class index of the image.

run_adversarial_attack(self)

Abstract method, its implementation will run the adversarial attack.

is_perturbed(self)

Abstract method, its method will return a boolean indicating whether a successful untargeted adversarial perturbation was found.

abstract is_perturbed()

The implementation of this method will return whether the adversarial attack has been successful

abstract run_adversarial_attack()

The implementation of this method will run the actual adversarial attack